As cloud technology is rapidly adopted and solutions are implemented for all types of cloud functions, the cost models for these solutions shift on a daily basis.
Just last week Amazon again lowered their pricing models for the S3 storage solutions within the cloud. All together, Amazon has lowered their prices 17 times for its cloud solutions in the last five years. Additionally, on February 15, Microsoft Azure announced a new pricing model that will potentially save customers 48-75 percent for databases larger than 1GB.
Why is the pricing model so variable for the cloud? Cloud providers such as Amazon, Rackspace and Azure want to better utilize their resources through the addition of “package” models and the changing landscape puts pressures to continue to cut the base costs. These models enable greater technology penetration in all types/levels of companies and competitive advantages for vendors.
For cloud users, this approach means they typically start optimizing their cloud environment by choosing “special offer” or attractive price model – and only then looking at the required capacity for their business. This approach often leads to the selection of the wrong pricing model. Businesses can’t select the right pricing model before they really understand what they will need in the cloud. This would be analogous to a consumer switching to another wireless provider because it is “cheaper” without analyzing his phone (calls, data, SMSs) historical usage.
Practically speaking, let’s assume that a cloud customer uses large instances in on-demand pricing model. The cloud vendor reached out to him with special offer to switch his pricing model to reservation model. The customer then switched to reservation pricing and started to look at his large instances utilization. He quickly realizes that these servers utilization is so low that he could use small instances and not large. However, because he has already reserved the large instances and he is over-provisioned and at a loss of how to rightsize.
Sophisticated algorithms are needed to find the most cost-effective provisioning options in real time resulting in increasing numbers of businesses coming to the realization that they don’t want or can’t implement management solutions on their own. Forecasting of actual usage – modeling of whether small or large instances or reservation models meet the needs of the business, capacity vs. performance, etc. – is critical to making the right pricing selection.
At Cloudyn, we’ve seen that with an analysis of the various models for pricing, businesses can see ~17% in savings with cloud price optimization alone. Together with rightsizing recommendations, the savings can increase to 40%+ of overall cloud costs — seen over and over with results from our customers.
The confusion and need for more insight in the cloud was evident to us with the response to our February 6th GA launch of the Cloudyn solution for cloud cost optimization. Since launching, we’ve seen overwhelming interest from companies thirsting for insight and guidance.
Over the next few weeks we’ll look at the technical aspects of working with pricing models and give examples of customer implementations benefiting from cost optimization.